CN109673015A - A kind of tracing area planing method based on spectral clustering - Google Patents
A kind of tracing area planing method based on spectral clustering Download PDFInfo
- Publication number
- CN109673015A CN109673015A CN201910000878.8A CN201910000878A CN109673015A CN 109673015 A CN109673015 A CN 109673015A CN 201910000878 A CN201910000878 A CN 201910000878A CN 109673015 A CN109673015 A CN 109673015A
- Authority
- CN
- China
- Prior art keywords
- network
- matrix
- subgraph
- cell
- indicates
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W16/00—Network planning, e.g. coverage or traffic planning tools; Network deployment, e.g. resource partitioning or cells structures
- H04W16/18—Network planning tools
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W16/00—Network planning, e.g. coverage or traffic planning tools; Network deployment, e.g. resource partitioning or cells structures
- H04W16/22—Traffic simulation tools or models
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Mobile Radio Communication Systems (AREA)
Abstract
The present invention proposes a kind of tracing area planing method based on spectral clustering, and this method is primarily directed to the usertracking regional planning method under the small cell network environment of hot spot region.Method Integral Thought proposed by the present invention is to be primarily based on the cellulor deployment of poisson process to construct a system model, generates display user mobility and pages the cellular network figure of characteristic；TA planning problem is then modeled as to the segmentation problem of figure related coefficient；Then spectral clustering of the application based on graph theory carries out TA planning to the network of generation.
Description
Technical field
The present invention relates to the user location administrative skill fields under the communications field, are directed primarily to a kind of pair of tracking area
The planing method of (Tracking Area, TA).
Background technique
In recent years, the rapid growth of mobile subscriber's quantity causes mobile communications network flow rapidly to increase, to the appearance of network
Amount requires to greatly increase.Large-scale hot spot region, such as there is a large amount of connection equipment in market.The intensive portion of small cell base station
Affix one's name to extremely urgent, super-intensive networking is come into being.Following small cell network disposes the hot spot region within the scope of macrocellular
Intensive self-organizing, low cost, the cellulor of low-power.Although small cell network has many advantages, due to cellulor
There is the features such as dynamic random deployment, super-intensive, self-organizing and strong liberalization ability, so that the mobile use under its coverage area
Family may continually migration and different cells, increase the complexity of user location management.
In the lte networks, macrocellular (eNB) is Chong Die with the coverage that cellulor (HeNB) covers.ENB passes through moving tube
Reason entity and gateway are connected to core network (Core Network, CN), and HeNB is correlated by the service network of Home eNodeB
It is connected to CN.Cell in network is divided into TA, and each TA has the unique identification code broadcasted by eNB, Yong Huke
Which TA be presently in identification.When user is moved to another TA from a TA, user is by the identification code of cell where it
It reports and gives home subscriber server (Home Subscriber Server, HSS), CN is inquired where called party by HSS
TAI, indicate TA in all base station call users.Therefore in the lte networks, in order to preferably carry out pipe to user location
Reason, the overlay area of network are divided into multiple TA, and each TA includes a large amount of cell, cannot be overlapped between TA.It is any logical
The non-occupied terminal execution position for crossing the boundary TA updates operation, and when there is calling to reach, CN will be to the affiliated TA model of user current location
All cells in enclosing send paging message.If each cellulor is individually divided into a TA, mobile subscriber will
It frequently executes location update operations and generates a large amount of location updating signaling.If all cellulors are divided into same
TA can generate huge paging cost when there is system calling arrival.Therefore, the purpose of TA planning is exactly using location updating as mesh
Mark solves the maximum boundary TA under paging capacity constraint, finds location updating cost and pages the equalization point between cost.
In order to more efficiently carry out TA planning to small cell network, the total signaling cost overhead of system, many TA rule are reduced
Cost-effective method is suggested.FU in 2013 et al. proposes a kind of post registration algorithm to reduce the signaling overheads in network, but should
Algorithm reduces system signaling expense to reduce flow relieving capacity as cost.Toril in 2013 et al. has counted mobile subscriber
In the moving characteristic of TA planning region, TA planning problem is modeled as figure segmentation problem, TA rule are provided using evolutionary search algorithm
The scheme of drawing, the algorithm belong to newest heuristic TA planning algorithm, and still, the algorithm steps are complicated, search TA planning problem
Globally optimal solution speed is slower, in the cellulor deployed environment for facing super-intensive networking, it is also necessary to the more efficient TA rule of research
Cost-effective method.Chen in 2017 et al. models to solve tracing area planning problem using evolution multi-objective Algorithm, which uses
The multiple-objection optimization of belt restraining, it is intended to find and preferably weigh between location updating signaling and call signaling.Ning in 2017 etc.
People proposes that TA planning problem is modeled as community in complex network and detected by a kind of TA planning algorithm based on community's detection, the algorithm
Problem provides TA programme using community's detection algorithm based on cooperation game, but when the cellulor in scene is more,
It will lead to and the un-reasonable phenomenon that more and more cellulors are separately divided into a TA occur.In addition, current most of dynamics
The research of position management method is all based on the same size of honeycomb, similar shape, with the regular cellular topology model of distribution, can not cope with
The cellular deployment environment of the super-intensive networking of hot spot region.In addition, with the variation that user's mobile trend and cellular base station are disposed,
Initial TA planing method cannot preferably optimize the signaling cost overhead of network system.
Summary of the invention
In view of the above-mentioned problems existing in the prior art, the present invention proposes a kind of tracing area planning side based on spectral clustering
Method, this method is primarily directed to the usertracking regional planning method under the small cell network environment of hot spot region.The present invention proposes
Method Integral Thought be to be primarily based on the cellulor deployment of poisson process to construct a system model, generate display and use
The cellular network figure of family mobility and paging characteristic；TA planning problem is then modeled as to the segmentation problem of figure related coefficient；
Then spectral clustering of the application based on graph theory carries out TA planning to the network of generation.
A kind of tracing area planing method based on spectral clustering mainly includes following implemented step:
Step 1, building generates the system model of cellular network figure
The present invention has built a system model first, and the small cell network system for counting hot spot region produces in for a period of time
Raw user's switching and paging data, building display user mobility and the small cell network figure for paging characteristic.Specific method is such as
Under.
(1) cellular deployment model
In order to reflect hot spot region small cell base station deployment randomness, the present invention use the Poisson based on random geometry
The cellular deployment model of point process.
(2) user distribution model
The present invention assumes user's Statistical Distribution under small cell base station to become dimensional gaussian distribution.
(3) user's mobility model
The present invention uses user's mobility model of random walk.
(4) system calling model
System calling model describes the frequency of user's calling and the duration of per call.The present invention uses Poisson mistake
The system calling model of journey.Assuming that calling arrival time t obeys the Poisson distribution of parameter μ, μ is that system calling reaches rate, general
Rate Density functional calculations formula are as follows:
F (t)=μ e-μt (1)
By the property of Poisson process it is found that the time interval obedience parameter that calling reaches isExponential distribution, then by mooring
The independent increment property of loose process generates.
Step 2, based on the TA planning problem modeling of figure segmentation
Reflection user mobility is generated by system model and pages the cellular network figure of characteristic, carries out the TA based on figure segmentation
Planning modeling.
(1) network G=(V, E) indicates to carry out the network of TA planning, wherein vertex set V={ v1,v2,...,vn}
And n × n matrix E respectively indicates syntople between cell and cell in network, wherein i is cell viIt numbers in a network,
N is number of cells in network.
(2) set P={ p1,p2,...,pnIndicate network in vertex weighted value, n be network in number of cells.Vertex
viWeight pi∈ P is cell viThe subscriber paging request number of times of middle generation.The weight e on sideij∈ E is cell viAnd vjBetween send out
Raw user's switching times, i and j are respectively cell viWith cell vjNumber in a network.
(3) by network graph partitioning at k regional ensemble C={ c1,c2,...,ck, small cell network is divided into not by expression
Same TA, wherein k indicates TA quantity.cx∈ C indicates a TA, and x is tracking area cxNumber in a network.Paging capacity B indicates one
In a TA, the maximum paging load of network support.
Then TA planning problem models are as follows:
Formula (2) indicates to minimize the mobile location updating generated of user, formula (3) table between the cell in different TA
Show the paging load summation of any one TA no more than paging capacity B.Next the spectral clustering based on graph theory will be used
Carry out TA planning.
Step 3, the TA planing method based on spectral clustering
Firstly, the L matrix of building network, then provides and cuts figure objective function, and by the property and optimization aim of L matrix
Function connects, and provides the optimization object function of TA planning.
(1) L matrix is constructed for network G=(V, E) in step 2, formula indicates are as follows:
L=D-E (4)
Wherein, matrix E is the adjacency matrix of network, and formula indicates are as follows:
Wherein, eij∈ E is defined as cell viWith cell vjBetween user's switching request number for occurring, i and j are respectively small
Area viWith cell vjNumber in a network, n are number of cells in network.
Matrix D formula indicates in formula (4) are as follows:
Wherein
(2) L matrix has a critical nature: L has following formula establishment for any one vector f:
(3) after the property for obtaining L matrix, building is undirected to cut figure
For the figure of cutting of non-directed graph G=(V, E), target is that figure is cut into the k subgraph not connected mutually, i.e. C=
{c1,c2,...,ck, any subgraph cx∈ C and cy∈ C meetsAnd c1∪c2∪...∪ck=V.Wherein x and y
Respectively subgraph cxWith subgraph cyNumber in non-directed graph.
For the set of any two subgraph pointcx,cyBetween cut figure weight are as follows:
So for k sub- set of graphs: C={ c1,c2,...,cK, figure function cut is cut in definition are as follows:
Wherein,Indicate cxSupplementary set,Indicate subgraph cxWith subgraphBetween all sides weight it
With this function optimization target is consistent with the minimum target of formula (2).But may be ineffective when cutting figure, therefore need
The scale restriction to each subgraph is limited, this is as the target of the constraint function of formula (3).TA further will be planned into mould
Type, which is converted into, cuts figure optimization aim, provides how to be defined to subgraph scale further below.
(4) figure optimization aim is cut
Use set P={ p1,p2,...,pnIndicate network in vertex weighted value, n be figure in number of cells.Vertex vi
Weight pi∈ P is cell viThe subscriber paging request number of times of middle generation.So for a subset c of Vx∈ C, is defined as:
Mostly not necessarily weight is just big for subgraph number of vertices, also more meets the target of TA planning when cutting figure based on weight, therefore
Improved objective function are as follows:
Wherein vol (cx) indicate subgraph cxIn include vertex weights and.
(5) in order to connect the property of L matrix with the objective function for cutting figure, definition instruction vector:
hx={ h1, h2..., hkX=1,2 ..., k
Wherein, k is the subgraph quantity for cutting figure, and x indicates subgraph cxNumber in figure.For any one vector hx, it is
One n-dimensional vector (n indicates number of vertex in figure), formula indicates are as follows:
Wherein, vi∈cxIndicate vertex viIn subgraph cxIn it is (same to indicate cell viIn tracking area cxIn),Indicate top
Point viNot in subgraph cxIn, i indicates vertex viNumber in figure, x indicate subgraph cxCutting the number in figure.
By the property of above-mentioned L matrix it is found that forHave:
Wherein, i and j respectively indicate vertex viAnd vertex viIt is numbered in figure.By formula (13) by the objective function of optimization with
The property of L matrix connects.For some subgraph cx∈ C, its NCut are corresponded toThen k subgraph corresponds to:
Wherein, matrix H is the matrix of k instruction vector composition.tr(HTLH) representing matrix HTThe mark of LH.Then cut figure optimization
Target are as follows:
L matrix is subjected to Eigenvalues Decomposition, the corresponding feature vector of k characteristic value before taking out, constituting size is n × k's
Eigenvectors matrix H, wherein n represents number of vertex in figure, and k represents the quantity of Qie Tu.Then primary simple k- is carried out to matrix H
Means algorithm cluster, k is also the clusters number of algorithm, and every a line represents a sample point.It obtains k and cuts figure C={ c1,
c2,...,ck, also correspond to k TA.
(6) the TA planning algorithm specific steps based on spectral clustering are given below:
1) the adjacency matrix E and matrix D for generating figure G=(V, E), construct L matrix；
2) wherein k is TA quantity to initialization k=1, and n is network cell number of cells；
3) circulation executes following steps, until k=n:
3.1 pairs of L matrixes carry out characteristic value solution, constitute the eigenvectors matrix H of n × k；
3.2 pairs of matrix Hs carry out primary simple k-Means algorithm cluster, clusters number k；
3.3 obtain k tracking area TA；
3.4k=k+1；
4) finally from the division of n times, take out so that the total signaling overheads of cellular network it is minimum divide as final TA
Program results.
Creativeness of the invention is mainly reflected in:
(1) present invention constructs a system model, counts in the small cell network system specific time of hot spot region and produces
Raw user's switching and paging data, the small cell network figure of building display user mobility and service switchover.
(2) present invention fully considers user mobility and paging characteristic in tracing area, using the spectral clustering based on graph theory
Algorithm provides TA program results.
(3) the system position turnover rate of the invention for being compared with other methods algorithm and total signaling overheads are relatively low.
Detailed description of the invention
Fig. 1 is present system model structure
Fig. 2 is present system model exemplary graph
Fig. 3 is that the present invention is based on the TA planning problem modeling figures of figure segmentation
Specific embodiment
Invention is described further below in conjunction with attached drawing and example.
Step 1, building generates the system model of cellular network figure
(1) variation for the randomness and user's mobile trend disposed for the cellulor of more preferable simulation hot spot region, this
Invention obeys one system model of poisson process model buildings, the main modular that system model is related to based on cellulor deployment
Structure is as shown in Figure 1.
(2) reflection is generated according to cellular deployment model, user distribution model, user's mobility model and system calling model to use
The network of family mobility and paging characteristic.As shown in Figure 2, wherein small cell base station distribution density λ=100, user's two dimension are high
The variances sigma of this distribution2=500,288 users are mobile around base station under the hot spot region 1000 × 1000m with speed v=1m/s
The track schematic diagram of 500 steps.
(3) by statistics a period of time in system model generate switching and paging data obtain display user mobility and
The small cell network figure of service switchover, finally TA planning algorithm of the application based on spectral clustering carries out TA division.
Step 2, based on the TA planning problem modeling of figure segmentation
(1) as shown in figure 3, network G=(V, E), which is indicated, will carry out the network of TA planning, wherein vertex set V={ v1,
v2,v3,v4,v5,v6And 6 × 6 matrix E respectively indicate syntople between cell and cell in network.
(2) set P={ p1,p2,p3,p4,p5,p6The weighted value that indicates vertex in network, number of cells is 6 in network.
Such as top cell v1Weight p1∈ P is cell v1The subscriber paging request number of times of middle generation.The weight e on side12∈ E is cell v1
And v2Between user's switching times (location update operations are carried out if across TA) for occurring.
(3) by Fig. 3 network graph partitioning at 2 regional ensemble C={ c1,c2, it indicates small cell network being divided into 2 not
Same TA.c1∈ C indicates a TA, wherein including cell v1, v5, v6。
TA planning problem can be then modeled as to formula (2), (3).One is connect to get off using the TA rule based on spectral clustering
The method of drawing carries out division network.
Step 3, the TA planing method based on spectral clustering
1) the adjacency matrix E and matrix D for generating figure G=(V, E), construct L matrix；
2) k=1 is initialized, and repeats step 3), 4), 5), until k=n.Wherein k is TA quantity, and n is that network cell is small
Area's quantity:
3) characteristic value solution is carried out to L matrix, constitutes the eigenvectors matrix H of n × k；
4) primary simple k-Means algorithm cluster, clusters number k are carried out to matrix H；
5) k tracking area TA is obtained；
6) finally from the division of n times, take out so that the total signaling overheads of cellular network it is minimum divide as final TA
Program results.
Claims (1)
1. a kind of tracing area planing method based on spectral clustering, which comprises the following steps:
Step 1, building generates the system model of cellular network figure
(1) cellular deployment model
Using the cellular deployment model of the poisson process based on random geometry；
(2) user distribution model
User's Statistical Distribution under small cell base station is assumed to become dimensional gaussian distribution；
(3) user's mobility model
Using user's mobility model of random walk；
(4) system calling model
System calling model describes the frequency of user's calling and the duration of per call；It is exhaled using the system of Poisson process
It is model；Assuming that calling arrival time t obeys the Poisson distribution of parameter μ, μ is that system calling reaches rate, probability density function
Calculation formula are as follows:
F (t)=μ e-μt (1)
By the property of Poisson process it is found that the time interval obedience parameter that calling reaches isExponential distribution, then by Poisson process
Independent increment property generate；
Step 2, based on the TA planning problem modeling of figure segmentation
Reflection user mobility is generated by system model and pages the cellular network figure of characteristic, carries out the TA planning based on figure segmentation
Modeling；
(1) network G=(V, E) indicates to carry out the network of TA planning, wherein vertex set V={ v1,v2,...,vnAnd n ×
N matrix E respectively indicates syntople between cell and cell in network, wherein i is cell viIt numbers in a network, n is net
Number of cells in network；
(2) set P={ p1,p2,...,pnIndicate network in vertex weighted value, n be network in number of cells；Vertex viPower
Weight pi∈ P is cell viThe subscriber paging request number of times of middle generation；The weight e on sideij∈ E is cell viAnd vjBetween the use that occurs
Family switching times, i and j are respectively cell viWith cell vjNumber in a network；
(3) by network graph partitioning at k regional ensemble C={ c1,c2,...,ck, it indicates small cell network being divided into difference
TA, wherein k indicates TA quantity；cx∈ C indicates a TA, and x is tracking area cxNumber in a network；Paging capacity B indicates one
In TA, the maximum paging load of network support；
Then TA planning problem models are as follows:
Formula (2) indicates to minimize the mobile location updating generated of user between the cell in different TA, and formula (3) indicates to appoint
Anticipate a TA paging load summation no more than paging capacity B；Next it will be carried out using the spectral clustering based on graph theory
TA planning；
Step 3, the TA planing method based on spectral clustering
Firstly, the L matrix of building network, then provides and cuts figure objective function, and by the property and optimization object function of L matrix
It connects, provides the optimization object function of TA planning；
(1) L matrix is constructed for network G=(V, E) in step 2, formula indicates are as follows:
L=D-E (4)
Wherein, matrix E is the adjacency matrix of network, and formula indicates are as follows:
Wherein, eij∈ E is defined as cell viWith cell vjBetween user's switching request number for occurring, i and j are respectively cell vi
With cell vjNumber in a network, n are number of cells in network；
Matrix D formula indicates in formula (4) are as follows:
(2) L matrix has a critical nature: L has following formula establishment for any one vector f:
(3) after the property for obtaining L matrix, building is undirected to cut figure
For the figure of cutting of non-directed graph G=(V, E), target is that figure is cut into the k subgraph not connected mutually, i.e. C={ c1,
c2,...,ck, any subgraph cx∈ C and cy∈ C meetsAnd c1∪c2∪...∪ck=V；Wherein x and y difference
For subgraph cxWith subgraph cyNumber in non-directed graph；
For the set of any two subgraph pointcx,cyBetween cut figure weight are as follows:
So for k sub- set of graphs: C={ c1,c2,...,cK, figure function cut is cut in definition are as follows:
Wherein,Indicate cxSupplementary set,Indicate subgraph cxWith subgraphBetween all sides the sum of weight, this letter
Number optimization aim is consistent with the minimum target of formula (2)；It but may be ineffective when cutting figure, it is therefore desirable to restriction pair
The scale of each subgraph limits, this is as the target of the constraint function of formula (3)；Further TA plan model will be converted to
Figure optimization aim is cut, provides how subgraph scale is defined further below；
(4) figure optimization aim is cut
Use set P={ p1,p2,...,pnIndicate network in vertex weighted value, n be figure in number of cells；Vertex viPower
Weight pi∈ P is cell viThe subscriber paging request number of times of middle generation；So for a subset c of Vx∈ C, is defined as:
Mostly not necessarily weight is just big for subgraph number of vertices, also more meets the target of TA planning when cutting figure based on weight, therefore improve
Objective function afterwards are as follows:
Wherein vol (cx) indicate subgraph cxIn include vertex weights and；
(5) in order to connect the property of L matrix with the objective function for cutting figure, definition instruction vector:
hx={ h1, h2..., hkX=1,2 ..., k
Wherein, k is the subgraph quantity for cutting figure, and x indicates subgraph cxNumber in figure；For any one vector hx, it is one
N-dimensional vector, n indicate number of vertex in figure, and formula indicates are as follows:
Wherein, vi∈cxIndicate vertex viIn subgraph cxIn, it is same to indicate cell viIn tracking area cxIn,Indicate vertex vi
Not in subgraph cxIn, i indicates vertex viNumber in figure, x indicate subgraph cxCutting the number in figure；
By the property of above-mentioned L matrix it is found that forHave:
Wherein, i and j respectively indicate vertex viAnd vertex viIt is numbered in figure；By formula (13) by the objective function of optimization and L square
The property of battle array connects；For some subgraph cx∈ C, its NCut are corresponded toThen k subgraph corresponds to:
Wherein, matrix H is the matrix of k instruction vector composition；tr(HTLH) representing matrix HTThe mark of L H；Then cut figure optimization aim
Are as follows:
L matrix is subjected to Eigenvalues Decomposition, the corresponding feature vector of k characteristic value before taking out constitutes the feature that size is n × k
Vector matrix H, wherein n represents number of vertex in figure, and k represents the quantity of Qie Tu；Then primary simple k- is carried out to matrix H
Means algorithm cluster, k is also the clusters number of algorithm, and every a line represents a sample point；It obtains k and cuts figure C={ c1,
c2,...,ck, also correspond to k TA；
(6) the TA planning algorithm specific steps based on spectral clustering are given below:
1) the adjacency matrix E and matrix D for generating figure G=(V, E), construct L matrix；
2) wherein k is TA quantity to initialization k=1, and n is network cell number of cells；
3) circulation executes following steps, until k=n:
3.1) characteristic value solution is carried out to L matrix, constitutes the eigenvectors matrix H of n × k；
3.2) primary simple k-Means algorithm cluster, clusters number k are carried out to matrix H；
3.3) k tracking area TA is obtained；
3.4) k=k+1；
4) it finally from the division of n times, takes out so that minimum the dividing of the total signaling overheads of cellular network is planned as final TA
As a result.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910000878.8A CN109673015B (en) | 2019-01-02 | 2019-01-02 | Tracking area planning method based on spectral clustering |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910000878.8A CN109673015B (en) | 2019-01-02 | 2019-01-02 | Tracking area planning method based on spectral clustering |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109673015A true CN109673015A (en) | 2019-04-23 |
| CN109673015B CN109673015B (en) | 2022-04-05 |
Family
ID=66146691
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910000878.8A Active CN109673015B (en) | 2019-01-02 | 2019-01-02 | Tracking area planning method based on spectral clustering |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109673015B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113269467A (en) * | 2021-06-15 | 2021-08-17 | 中国电信股份有限公司 | Region planning method and device based on graph segmentation, storage medium and electronic equipment |
| CN113543280A (en) * | 2021-05-27 | 2021-10-22 | 新华三技术有限公司成都分公司 | Neighbor relation discovery method and device, electronic equipment and storage medium |
| CN114173363A (en) * | 2022-02-10 | 2022-03-11 | 深圳市优网科技有限公司 | Method and device for detecting TAC planning abnormity and storage medium |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105323769A (en) * | 2014-06-12 | 2016-02-10 | ***通信集团贵州有限公司 | Tracking region planning method and system |
| CN105636065A (en) * | 2014-10-31 | 2016-06-01 | ***通信集团公司 | TA (tracking area) planning method and equipment for nano base station |
| CN106851664A (en) * | 2015-12-03 | 2017-06-13 | 中国联合网络通信集团有限公司 | Determine the method and device on tracking area border |
| CN107808518A (en) * | 2017-10-26 | 2018-03-16 | 东南大学 | A kind of traffic zone division methods theoretical based on multi-path spectral clustering |
| US20180165302A1 (en) * | 2016-12-12 | 2018-06-14 | Facebook, Inc. | Systems and methods to provide local suggestions based on spectral clustering |
-
2019
- 2019-01-02 CN CN201910000878.8A patent/CN109673015B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105323769A (en) * | 2014-06-12 | 2016-02-10 | ***通信集团贵州有限公司 | Tracking region planning method and system |
| CN105636065A (en) * | 2014-10-31 | 2016-06-01 | ***通信集团公司 | TA (tracking area) planning method and equipment for nano base station |
| CN106851664A (en) * | 2015-12-03 | 2017-06-13 | 中国联合网络通信集团有限公司 | Determine the method and device on tracking area border |
| US20180165302A1 (en) * | 2016-12-12 | 2018-06-14 | Facebook, Inc. | Systems and methods to provide local suggestions based on spectral clustering |
| CN107808518A (en) * | 2017-10-26 | 2018-03-16 | 东南大学 | A kind of traffic zone division methods theoretical based on multi-path spectral clustering |
Non-Patent Citations (3)
| Title |
|---|
| LEI NING等: "Tracking Areas Planning with Cooperative Game in Heterogeneous and Small Cell Networks", 《INTERNATIONAL CONFERENCE ON GAME THEORY FOR NETWORKS》 * |
| MATÍAS TORIL等: "Automatic Replanning of Tracking Areas in Cellular Networks", 《IEEE TRANSACTIONS ON VEHICULAR TECHNOLOGY 》 * |
| 李芸: "LTE网络跟踪区优化技术研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113543280A (en) * | 2021-05-27 | 2021-10-22 | 新华三技术有限公司成都分公司 | Neighbor relation discovery method and device, electronic equipment and storage medium |
| CN113269467A (en) * | 2021-06-15 | 2021-08-17 | 中国电信股份有限公司 | Region planning method and device based on graph segmentation, storage medium and electronic equipment |
| CN114173363A (en) * | 2022-02-10 | 2022-03-11 | 深圳市优网科技有限公司 | Method and device for detecting TAC planning abnormity and storage medium |
| CN114173363B (en) * | 2022-02-10 | 2022-04-15 | 深圳市优网科技有限公司 | Method and device for detecting TAC planning abnormity and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109673015B (en) | 2022-04-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| JP6320939B2 (en) | An evolutionary algorithm for geographical load balancing using distributed antenna systems | |
| Tsai et al. | Metaheuristics for the deployment of 5G | |
| CN109673015A (en) | A kind of tracing area planing method based on spectral clustering | |
| CN101547450B (en) | Method and apparatus for zoning in mobile communication network | |
| Vidyarthi et al. | A hybrid channel assignment approach using an efficient evolutionary strategy in wireless mobile networks | |
| CN110430579B (en) | Wireless AP deployment optimization method based on fruit fly optimization and used in non-uniform environment | |
| CN103810270B (en) | Optimized tracking area partition method and device | |
| CN106358300B (en) | A kind of distributed resource allocation method in microcellulor network | |
| CN109831790B (en) | Cooperative caching method based on brainstorming optimization algorithm in fog wireless access network | |
| CN107659973A (en) | Super-intensive network cluster dividing method based on density K means algorithms | |
| CN110366187B (en) | Wireless network base station building site selection optimization method for balancing cost and signal intensity | |
| Chen et al. | Modeling the tracking area planning problem using an evolutionary multi-objective algorithm | |
| CN113676919B (en) | Network coverage enhancement method based on resource cell self-adaptive generation | |
| CN108063802A (en) | User location dynamic modeling optimization method based on edge calculations | |
| CN109600756A (en) | A kind of physical area identification distribution method based on the preferential coloring algorithm of maximal degree | |
| CN106658526A (en) | Simulated annealing algorithm based frequency spectrum distribution method in super-dense small cell network | |
| CN109548138A (en) | A kind of tracking area list managing method based on the detection of overlapping community in small cell network | |
| CN109451507A (en) | A kind of tracing area planing method based on community's detection in isomery cellular network | |
| CN109831794B (en) | Base station clustering method based on density and minimum distance in ultra-dense network | |
| CN106817691A (en) | A kind of PCI distribution method and device | |
| Alenoghena et al. | Application of genetic algorithm in radio network coverage optimization–a review | |
| CN102026239A (en) | Frequency optimization method based on measurement report | |
| CN108566645A (en) | Network user's cut-in method based on Neyman-Scott cluster processes | |
| CN110380808B (en) | Micro-cell semi-clustering interference coordination method taking user equipment as center | |
| CN103987052B (en) | Base station collocation method and base station, telegon |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |